US 7833897 B2
A method is provided for making of interconnect solder bumps on a wafer or other electronic device without depositing any significant amount of tin or other solder component from the solder onto the wafer surface which tin can cause shorts or other defects in the wafer. The method is particularly useful for well-known C4NP interconnect technology. In one aspect of the invention, a reducing gas flow rate is used to remove oxides from the solder surfaces and wafer pad surfaces and is of a sufficient determined or pre-determined flow and/or chamber or mold/wafer spacing to provide a gas velocity across the solder surfaces and wafer pad surfaces so that Sn or other contaminants do not deposit on the wafer surface during solder transfer. In another aspect, the transfer contact is performed below the melting point of the solder and subsequently heated to above the melting temperature while in transfer contact. The heated solder in contact with the wafer pads is transferred to the wafer pads.
1. A method for forming solder bumps on an electronic substrate surface comprising the steps of:
providing a solder transfer mold having a plurality of solder mold openings in an array, which mold openings are filled with solder;
positioning the solder containing openings of the mold proximate the surface of an electronic substrate to which the solder is to be transferred, the array openings in the mold corresponding to an array on the electronic substrate surface;
providing a seal around the mold and the electronic substrate and forming a chamber;
heating the chamber to an elevated temperature preferably sufficient to liquify the solder;
passing a reducing gas into the chamber contacting the mold and the electronic substrate to remove oxides from the mold solder and electronic component surfaces and removing the gas after passing through the chamber;
determining the flow rate of the reducing gas into the chamber needed to provide a sufficient gas velocity to prevent any significant Sn or other solder components from depositing on the substrate surface;
maintaining the determined flow rate during the transfer of the solder from the mold to the substrate;
transferring the solder from the mold openings to the corresponding array on the electronic substrate surface by bringing the solder mold and electronic substrate into transfer contact or near contact; and
removing the seal from the around the mold and removing the solderless mold from the electronic substrate leaving the electronic substrate with the solder from the mold openings on the surface thereof.
2. The method of
3. The method of
4. The method of
5. The method of
6. The method of
1. Field of the Invention
The present invention relates to the making of interconnect solder bumps on a wafer or other electronic device without depositing any significant amount of tin from the solder onto the wafer surface which tin can cause shorts or other defects in the wafer, and, in particular, to the well-known C4NP interconnect technology.
2. Description of Related Art
Forming an electronic package assembly whereby an electrical component such as an integrated circuit chip is electrically and mechanically connected to a substrate, a card, or board, another chip or another electronic part is well-known in the art. This technology is generally termed surface mount technology (SMT) and has gained acceptance as the preferred means of making electronic package assemblies. The interconnect technology is commonly known as ball grid array packaging, C4 flip chip interconnect, multi-chip modules, multilayer and micro via printed wiring boards and surface mount hybrid assembly.
Multilayer ceramic and organic electronic components are typically joined to other components by soldering (or capture) pads on a surface of one of the electronic components to corresponding soldering (or capture) pads on the surface of the other component. Control Collapse Chip Connection is an interconnect technology developed by IBM as an alternative to wirebonding. This technology is generally known as C4 technology or flip chip packaging. Broadly stated, an integrated circuit chip is mounted above a ceramic or organic substrate and pads on the chip are electrically and mechanically connected to corresponding pads on the substrate by a plurality of electrical connections such as solder bumps to form an electronically connected module. A module is typically connected to other electronic components by solder or socket type connections.
In the C4 interconnect technology a relatively small solder bump is attached to the pads on one of the components being joined, typically to the chip. The electrical and mechanical interconnects are then formed by positioning the corresponding pads on the substrate to be joined adjacent the solder bumps on the chip and reflowing the bumps at an elevated temperature. The C4 joining process is self-aligning in that the wetting action of the solder will align the chip bump pattern to the corresponding pads on the substrate. Capture pads can also be sputtered and subsequently etched without any plating.
Capture pads for C4 bumps on semiconductor wafers or on the substrate to be interconnected are well-known and are typically made by a through resist plating of Ni or Cu/Ni pads onto a Cu seed layer. It is also preferred to use a conductive barrier layer on the substrate surface and the Cu seed layer is preferably made by sputtering of Cu onto a sputtered TiW layer.
In C4 technology the solder bumps are formed directly on the capture pads of the one unit. The pads are electrically isolated from other pads by the insulating chip passivation and substrate that surrounds each pad. The substrate may be un-doped silicon (Si) or some other material. The bottom of the pad is electrically connected into the chip or substrate circuit.
A major application of C4 is in joining semiconductor microchips (integrated circuits) to chip packages. Chips usually are made in rectangular arrays on a mono-crystalline slab of silicon called a “wafer,” which is a thin disc several inches across. Many chips are formed on each wafer, and then the wafer is diced into individual chips and the chips are “packaged” in units large enough to be handled. The C4 bumps are placed on the chips while they are still joined in a wafer.
One method of forming solder bumps uses sputtering or vacuum deposition. Solder metal is evaporated in a vacuum chamber and the metal vapor coats everything in the chamber with a thin film of the evaporated metal. To form solder bumps on the substrate, the vapor is allowed to pass through holes in a metal mask held over the substrate. The solder vapor passing through the holes condenses onto the cool surface into solder bumps. This method requires a high vacuum chamber to hold the substrate, mask, and flash evaporator.
An alternative technique for making solder bumps is electrodeposition, also called electrochemical plating or electroplating. This method also uses a mask such as patterned photoresist and forms solder bumps only at the selected sites. Other methods include screening a solder paste through a mask and depositing micro-bumps in place.
The C4NP transfer process is the preferred method and uses a mold containing the solder and transfers the solder in the mold to the capture pad of the wafer and upon heating forms solder bumps on the wafer pads. The process however may result in organo-Sn and Sn or other components of the solder being deposited on the surface of the wafer during transfer processing. These wafers will be rejected due to these surface deposits because the presence of these deposits is a source of high resistance shorts between C4's and paths for electro-migration fails during field use. The mold solder bumps are preferably filled with solder by an Injection Molding Solder (IMS) method as described below.
Bearing in mind the deficiencies of the prior art it is an object of the present invention to improve the C4NP process and in particular the method for the transfer of the solder in the mold to the capture pad of the wafer substrate.
Still other objects and advantages of the invention will in part be obvious and will in part be apparent from the specification.
The above and other objects, which will be apparent to those skilled in the art, are achieved in the present invention which is directed to a method for forming solder bumps on an electronic substrate surface comprising the steps of:
In another aspect of the invention a method is provided for forming solder bumps on an electronic substrate surface comprising the steps of:
The features of the invention are believed to be novel and the elements characteristic of the invention are set forth with particularity in the appended claims. The figures are for illustration purposes only and are not drawn to scale. The invention itself, however, both as to organization and method of operation, may best be understood by reference to the detailed description which follows taken in conjunction with the accompanying drawings in which:
In describing the preferred embodiment of the present invention, reference will be made herein to
Various solders may be readily processed using the method of the invention and these include lead and lead-free ternary and quaternary alloys. Lead-free alloys are of particular commercial significance and for convenience the following description will be directed to such lead-free alloys although it will be appreciated by those skilled in the art that any suitable solder may be readily processed using the method of the invention. Exemplary lead-free alloys include Sn-1.8% Ag, Sn-0.5% Ag and Sn-0.7% Cu. Also, in particular, the following description will be directed to tin containing lead-free alloys since these are likewise of significant commercial interest and an efficient method of providing solder interconnects with these alloys is of important commercial consideration.
IMS is the preferred method to make mold plates containing cavities which are filled with solder and a preferred transfer device for use in the method is shown in
The mold plates can be made from a variety of materials. Since transfer of the solder from the mold plate to the final solder receiving substrate, such as the wafer, occurs at elevated temperatures, matching the coefficient of thermal expansion (CTE) of the mold plate and substrate is important. This is especially true as the area to be transferred increases. Thus, for a single chip, or even a 4 inch diameter wafer, the smaller distance to a neutral point is such that the mold and substrate can tolerate some CTE mismatch and yet still works successfully. However, for larger areas such as 8 inch and 12 inch diameter wafers it is important that the mold material closely match the CTE of the wafer substrate, i.e., silicon.
Cavities (the solder openings) 14 in the mold plate are in a pattern and are the mirror image of the solder receiving pads on the final substrate or wafer. The cavities can be produced in a mold plate by any number of techniques, the selection of which is dependent upon the cavity size and pitch as well as the mold plate material. Cavity volume uniformity is essential since they directly determine the solder bump volume on the wafer.
The above method as shown in
Alignment of the mold and the transfer substrate, i.e., wafer, is critical to the success of wafer mold or bump transfer. When transparent mold plates are used, it is relatively easy to align solder filled cavities to wafer pads. When non-transparent material is used, or when a gaseous flux is used, alignment using split optics may be necessary. Depending upon the environment of the transfer, flux may or may not be used. If used, it can be applied in a thin even coat onto the filled mold plate or wafer before proceeding to the transfer fixture. Or it can be applied in the vapor phase at or near the solder transfer temperature. If not used, oxide reducing methods such as pressure variation or hydrogen reflow can help to facilitate the transfer. In general, when the solder in the mold plate is in a liquidus state and the wafer pads are oxide free, the solder wetting forces exceeds surface tension forces that maintain the molten solder in their cavities. After cooling to solidify the solder, the solder bumps are released from the mold plate onto the wafer pads and the mold plate is lifted from the wafer. The shape of the bumps on the wafer tend to be spherical, especially if the capture pad on which they are transferred is round. The wafer can now be independently subjected to a final solder reflow excursion after wafer test to obtain spherically shaped bumps on the wafer pads if test damage makes this necessary.
The present invention is directed to improving the above transfer process and it is believed that the source of tin deposits on the wafer surface are from the solder in the mold and that the deposits occur prior to contact of the solder in the mold to the capture pad on the wafer.
The Sn contamination problem is believed to arise when one of two mechanisms occurs. In the first mechanism, some of the tin from the solder in the mold is vaporized and redeposited onto the wafer surface. The vapor pressure of Sn is fairly low, however, in the presence of formic acid, it is much greater. The second possible mechanism is that the solder in the mold has an oxide coating that is thinned during the formic acid flow. When a pin hole opens in the solder oxide surface, the liquid solder under the oxide is under pressure due to the volumetric expansion during melting and it will “squirtout” through this pin hole and spray onto the wafer surface. The liquid solder may be atomized as it rapidly exits the mold cavity through this pin hole in the oxide surface and microscopic tin balls may be formed on the wafer surface.
The subject invention is directed to methods to eliminate the surface tin contamination of the wafer during the C4NP transfer process.
In a first aspect of the invention as shown in
It is an important aspect of the invention that any Sn forming as a vapor not be deposited on the wafer surface and this is accomplished by maintaining an effective gas flow rate across the wafer and pad surfaces which sweeps away any vaporized Sn or other solder components during the transfer process. It will be appreciated by those skilled in the art that the gas velocity across the mold surface and wafer surface is dependent on a number of parameters such as the gas flow rate in and out of the chamber, the space between the mold and the wafer surface and the size of the containment chamber. For example, for the same gas flow rate in and out of the chamber and the same space between the mold and the wafer, a larger size containment chamber will provide a lower gas velocity across the surface of the mold and wafer. Similarly, for the same size containment chamber and space between the mold and the wafer, increasing the gas flow rate in and out of the chamber will increase the velocity across the wafer surface and mold surface. Likewise, for the same size containment chamber and gas flow rate in and out of the chamber, increasing the space between the mold surface and the wafer surface will decrease the velocity across the mold surface and wafer surface. The effective gas flow rate for a particular transfer process can be determined empirically.
It has been found for a 12 inch wafer using as the seal a confinement chamber which is about 13 inch diameter and 2 mm high that a reducing gas flow rate of about 3.5 liter/minute was ineffective to prevent the redeposition of Sn on the wafer surface. Doubling the flow rate to about 7 liter/minute prevented the formation of Sn deposits on the wafer surface. In this example the space between the mold surface and wafer surface was the same and was about 2 mm. If the space was increased it would be preferable to likewise increase the gas flow rate into the chamber above the 7 liter/minute flow rate. If a smaller confinement chamber is used, a lower gas flow rate in and out of the chamber can be used to provide the sufficient gas velocity to prevent Sn contamination of the wafer surface.
Broadly stated, if, in the operation of the C4NP transfer process it is found that there is Sn or other solder component contamination on the wafer surface, the Sn contamination can be prevented following the method of the invention by either increasing the reducing gas flow rate into the chamber, reducing the size of the chamber at the same flow rate and/or reducing the space between the wafer surface and mold surface. If, for example, as noted above, the confinement chamber and space between the wafer and mold surfaces are kept constant, increasing the flow rate of the reducing gas will increase the gas velocity across the wafer surface and mold surface and prevent Sn deposition the wafer surface.
In a preferred aspect of the invention, an effective flow rate into the chamber is determined or pre-determined for a specific chamber, wafer size, and spacing and this or a higher flow rate is maintained during the transfer process. Deposition of Sn on the wafer surface can be determined by examination of the wafer visually. For many cases, 1× visual is adequate. Further confirmation can be done with a low powered microscope. XPS can be used to get a quantitative Sn concentration.
In a second aspect of the invention as shown in
Referring now to
The improved C4NP process of this invention is applicable to pad layers such as TiW/Ni, TiW/Co, TiW/Cu/NiP, TiW/NiV, TiW/NiV/Cu, TiW/NiSi, TiW/NiSi/Cu, and other Ni and Cu alloys that are compatible with the solder C4s. Further, while the above description describes the mold below the electronic substrate during transfer, for some methods the mold may be positioned above the electronic substrate.
While the present invention has been particularly described, in conjunction with a specific preferred embodiment, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art in light of the foregoing description. It is therefore contemplated that the appended claims will embrace any such alternatives, modifications and variations as falling within the true scope and spirit of the present invention.
Thus, having described the invention, what is claimed is: